views:

81

answers:

2

Being able to distinguish a GUID from random data can be useful when debugging obscure code defects.

On Windows each GUID generated is of version 4 therefore it has '4' as the first half-byte of the third part. So if the 16-byte sequence violtates that rule it is not a version 4 GUID. For example,

567E1ECB-EA1C-42D3-A3ED-87A5D824D167

could be either a version 4 GUID or anything else, but

567E1ECB-EA1C-02D3-A3ED-87A5D824D167 //third section starts with 0, not with 4

is not a version 4 GUID.

What are other signs of a 16-byte memory block being or not being a valid GUID?

+1  A: 

GUID's are hexadecimal, so you could check what characters are included (i.e. 'X', 'Y', etc. are not valid)

Check wikipedia for the definition and you could probably find out more ideas there.

http://en.wikipedia.org/wiki/Globally_Unique_Identifier

Stuart
In a 16 byte memory block, you have only bytes, not characters, and in its hexadecimal representation (32 nibbles) by definition you have none of those.
MSalters
+1  A: 

Besides GUID version (0100), there's also something called the GUID variant. This will be 2 bits (10) in octet 8. The remaining bits a of v4 GUID are by definition random.

MSalters
Do you mean that in every GUID the first 2 bits in octet 8 must be "10"?
sharptooth
Actually the variant is the three most significant bits of octet 8 - and can be `000`, `010`, `110` or `111`. The standard one is `010` though and for GUIDs generated in the same way these bits will always have the same value.
configurator
Sorry, I meant octet 10.
configurator